A six month, full time research internship, starting January 2020 is available at Oracle Labs

The objective of this internship is to detect program bugs through semantic modelling of source code. As bug detection is usually considered as a classification problem, it is crucial to select suitable features that can distinguish buggy programs from bug-free ones. However, identifying such discriminating features through manual analysis or heuristics still remain challenging.

This internship will explore Graph Neural Networks (GNN) to model source code such that program semantics are precisely captured. Graphs are the basic data-structure for any analysis of programs, hence, graph-based source code representation is ideal for bug detection. Programs are represented as abstract syntax tree, control-flow or data-flow graph and neural graph embedding technique is applied to learn a vector representation of programs. Finally, machine learning classifier (e.g, SVM) is used to detect whether a program is buggy or bug-free. Specifically, we are interested in detection of memory related bugs in C/C , such as use-after-free, double-free and so on.

Supervisors\: Mahin Chandramohan, Cristina Cifuentes

Mahin Chandramohan is a Senior Researcher for Oracle Labs, Australia, leading machine learning efforts in the Program Analysis group. He received his bachelor's and doctoral degree in Computer Science from Nanyang Technological University, Singapore. Mahin has extensive experience in the malware analysis, binary vulnerability auditing, program analysis and machine learning.

Cristina Cifuentes is the Senior Director of Oracle Labs Australia. She has incepted and transferred a variety of research and advanced development projects, including Parfait the bug tracking project. Her research interests include programming languages, compiler construction and software security.

About Oracle Labs\:

Oracle Labs, the research arm for Oracle, focuses on applied research that produces new technologies of interest to the company. Oracle Labs Australia, based in Brisbane, focuses on Program Analysis as it applies to a variety of domains, including bug-checking, security analysis, productivity tools, testing, and more. The group is best known for its research on static code analysis that led to scalable and precise bug-checking algorithms embedded in the Oracle Parfait tool. For more information, visit http\://labs.oracle.com/locations/australia

Oracle internships give students valuable industry experience and the chance to work on cutting-edge research projects with real-world applications.


Essential Criteria\:

You will be expected to\:
   * be undertaking a Master or PhD in Computer Science or related topic,
   * have a good understanding of program analysis concepts, especially program
   * representation techniques (e.g., abstract syntax tree (AST), control-flow  graph (CFG), etc.),
   * have a good background in machine learning,
   * be proficient in Python,
   * have a good understanding of C/C ,
   * demonstrated ability to work independently and collaboratively.

Good to have\:

   * Willingness to read and explore scientific literature in this field
   * Good understanding of Natural Language Processing (NLP) concepts and techniques
   * Hands-on experience with popular deep learning frameworks (e.g., Tensorflow, PyTorch, etc...)

Contact Paddy Krishnan paddy.krishnan@oracle.com for more details. The position will commence interviewing immediately and remain open until filled.
!|!This job code is utilized for the majority of our temporary hires. The individual is performing hourly job duties as defined under the Fair Labor Standards Act.



!|!